Education evaluation agencies in China face challenges in avoiding moral hazards when assessing higher education institutions. With multiple agencies involved, including official, semi-official, university-affiliated, and third-party agencies, the risk of moral hazard increases. By analyzing the situation using critical discourse analysis, it was found that implementing a "relative performance" evaluation mechanism can effectively mitigate moral hazards in higher education evaluation under the "multi-agent" mode. This approach not only helps prevent moral hazards but also enhances the overall effectiveness of the evaluation process.